Abstract
BACKGROUND: Brain metastasis (BM) is one of the common ways of tumor metastasis and has a poor prognosis. This study aims to identify potential biomarkers from the perspective of somatic mutations, providing a basis for the prognosis evaluation and immunogenicity prediction of BM patients. METHODS: This study collected the somatic mutation profiles and clinical information of a total of 421 patients with BM in Memorial Sloan Kettering Cancer Center (MSKCC). Non-negative matrix factorization was employed to extract the mutational process signatures operating in the genome. Consensus clustering analysis was utilized to identify mutation-related molecular subtypes. Through a comprehensive analysis of genomic mutations and copy number variations (CNV), biomarkers associated with outcomes and tumor immunogenicity were screened. RESULTS: Non-small cell lung cancer, melanoma, and breast cancer were common primary tumors of BM, and these three tumor types exhibited better prognosis compared to other types. This study found that a higher tumor mutation burden (TMB) was significantly associated with a better prognosis of BM. A total of four mutational process signatures were extracted, and among them, a signature featured by C > T mutations and related to DNA damage repair was proven to be linked with an inferior outcome and a lower TMB. Through integrated genomic mutation analysis, PTPRT mutation was determined to associate with improved prognosis of BM. More importantly, patients carrying this mutation also harbored a better response to immunotherapy. CNV analysis indicated that PTEN deletion and DUSP4 deletion were respectively associated with poorer and better outcomes in patients with BM. CONCLUSIONS: By integrating the somatic mutation data of patients with BM, this study identified molecular biomarkers related to outcomes and immunogenicity from three perspectives: mutational process signatures, molecular subtypes, and genomic variations. Our findings provide clues for prognosis evaluation in BM patients. They also establish a theoretical basis for predicting immunotherapy efficacy.